D. A. Dominguez1, S. Ivry3, S. Hatcher3, E. Gilbert2, S. Kumar2, W. Park6, M. Schmidt7, R. Brand4, A. O’Donoghue5, K. Kirkwood2, C. Craik3 1UCSF East Bay,Department Of General Surgery,Oakland, CA, USA 2UCSF,Department Of General Surgery,San Francisco, CA, USA 3UCSF,Department Of Pharmaceutical Chemistry,San Francisco, CA, USA 4University Of Pittsburgh,Department Of Gastroenterology,Pittsburgh, PA, USA 5UCSD,Department Of Pharmacy And Pharmaceutical Sciences,San Diego, CA, USA 6Stanford University,Department Of Gastroenterology,Palo Alto, CA, USA 7Indiana University,Department Of Surgery,Indianapolis, IN, USA
Introduction: Risk stratification of pancreatic cystic lesions remains an area of great clinical uncertainty. Measurement of cyst fluid CEA is used to predict which cysts are mucinous, and therefore may have malignant potential, but its accuracy limits its usefulness. Due to this lack of definitive pre-operative characterization, some patients will die from undiagnosed cancers, while others undergo unnecessary pancreatic resections with significant morbidity. Dysregulation of protease expression and activity has been previously reported in mucinous pancreatic cyst fluid. We tested the hypothesis that differences in proteolytic activity between mucinous (MUC) and non-mucinous (NON-MUC) cysts could be used to improve our ability to identify pre-malignant pancreatic tumors.
Methods: We first analyzed cyst fluid from a cohort of human pancreatic neoplasms (MUC, n=16; NON-MUC, n=7) using multiplex substrate-profiling by mass spectrometry (MSP-MS), which is an unbiased, comprehensive technology for analyzing patterns of proteolytic activity. We found that aspartyl protease activity was unique to MUC pancreatic cysts. Next, using shotgun proteomic analysis, we identified 2 proteases (Prot1, Prot2), that were selectively expressed in MUC cyst fluid. Specific cleavage profiles were used to design a selective fluorescent substrate for each protease. Substrates were then used to assess the activity of each protease in a larger patient cohort (MUC, n=71; NON-MUC, n=39); ROC curves were generated for each substrate. Performance was compared with CEA values (CLIA lab) using >= 192 ng/mL as a cut off for MUC cysts.
Results: Receiver operator characteristic (ROC) curves for Prot1 and Prot2 exhibited an area under the curve (AUC) of 0.98 and 0.82, respectively. Prot1 demonstrated a sensitivity of 93% and specificity of 100%. Prot2 had a sensitivity of 70% and a specificity of 92%. Testing required < 0.2 mL cyst fluid. CEA had an AUC of 0.86, and a sensitivity and specificity of 65% and 94%, respectively.
Conclusion: MSP-MS is a powerful technology to examine dysregulated proteolysis in complex biological fluids. The resultant fluorogenic substrates outperformed CEA, the highest of which had a sensitivity of 93% and a specificity of 100%, for the detection of MUC pancreatic cysts. Our fluorescent assay has the potential to be a rapid, inexpensive, and highly accurate predictor of MUC pancreatic cysts.
